[llamaimperative]

But you can implement transactions without any recourse trivially using existing technology today...

> Web3 can prevent a whole class of problems that normally would require after-the-fact arbitration to fix.

Such as?

> This is the massive improvement over code that runs inside some server farm and can be switched up anytime.

Massive theoretical improvement. No debate there. The question is about applicability, which I notice you've still not made concrete.

> It's weird arguing that no, we would rather have databases where certain people have the key to subtly corrupt the entire database

I don't know, it basically seems to work and as far as I can tell, you're about a billion times more likely to have your shit stolen from you in the web3/cryptoverse than using the traditional financial system... except of course if you're a criminal and/or a more grey-area "unsavory character" to some of the main stakeholders. Which that is a good usecase for this tech, of course.

[EGreg]

Are you serious? There are so many problems it can prevent, because they are nearly impossible to occur now:

1) People can only act as themselves, they can't log into the database and corrupt all the data in a table, so there is no need to recover from massive systemic corruption (this is the big one)

2) Business rules of smart contracts are enforced, and you know that the state transitions happen exactly as they should (this is the reason that databases use stored procedures, for example)

I can list a few, but there are literally thousands:

1) In an auction, the highest bidder is the one who actually wins, it can't be hacked. You can have N winners, and once the N+1st arrives, it returns the money to the lowest bidder who can bid again. Bids increase at a predictable rate, everyone knows the deal upfront, etc.

As opposed to, say, this on the highest level: https://observer.com/2018/03/christies-sothebys-lawsuits-sho...

2) Salaries can be paid out to specific addresses, at a specific rate, and be approved by managers, with the history of all this preserved

3) Atomic swaps between tokens, escrow, each side knows the money is there and will be released if they do their part

4) Rules about staking reserves, guarantee the money is there for vendors to cash out, vendors must be whitelisted to cash out, etc. etc.

5) Large Communities can enforce their "constitution" and which roles invite which roles, at what rate, etc. Ah here, just look:

https://community.intercoin.app/t/intercoin-defi-communityco...

and so on and so forth

[habinero]

> People can only act as themselves, they can't log into the database and corrupt all the data in a table, so there is no need to recover from massive systemic corruption

I don't understand why you think this is so critical? This is a solved problem - just have backups and basic security. If it's really critical, there's a lot of ways to log changes. It's not a big deal.

> Business rules of smart contracts are enforced, and you know that the state transitions happen exactly as they should

Ok, but we have that already. They're called regular contracts. If they have bugs or unexpected problems or someone doesn't follow through, you can get the state to make them.

> In an auction, the highest bidder is the one who actually wins, it can't be hacked

And how often is this a problem, exactly? The article says it happens maybe once every five or six years.

> 2-5

We have all of this already. Banks, HR software, escrow, reserves. It works fine.

You want the ability to unwind a contract or a transaction if something goes wrong. In fact, the contracts are only really used if things go wrong. Most of the time, the parties just follow the contact voluntarily.

[llamaimperative]

To add on: if someone's "irreversible" crypto transaction goes awry, do people really think they're not going to go to court and/or that courts won't do exactly what courts do in every other case of fraud?

"Your honor, the smart contract was clear as day: this fraud was totally above board" is not going to work. I promise.